Using machine learning techniques for age prediction based on PPG signal analysis

Proceedings of International Scientific Conference „ALFATECH – Smart Cities and modern technologies“ (pp. 154-159) 

 

AUTOR(I) / AUTHOR(S): Mirjana Tomić , Stevan Jokić, Ivan Jokić, Nenad Gligorić, Ana Kovačević, Branislav Gerazov 

 

Download Full Pdf   

DOI:  10.46793/ALFATECHproc25.154T

SAŽETAK / ABSTRACT:

This paper explores the application of machine learning and neural networks for age prediction based on PPG signals (photoplethysmographic signals), which provide a non-invasive and cost-effective method for assessing patient health, particularly in the field of cardiovascular diseases. PPG signals, recorded using light sources and photodetectors, enable the assessment of changes in blood volume in the microvasculature, which is closely related to the condition of blood vessels. In this study, a machine learning model based on neural networks has been developed, using PPG signals as input data to predict patient age.

Various neural network architectures were tested, including models with one hidden layer and models with multiple layers, to investigate how the number of layers affects prediction accuracy. Additionally, different activation functions, such as tanh and ReLU, as well as various data preprocessing techniques, such as normalizing PPG signals, were considered. The model evaluation was carried out using MAE (Mean Absolute Error) and MSE (Mean Squared Error) as key statistical indicators measuring prediction accuracy.

The results show that models with a greater number of hidden layers achieve better performance in age prediction, with a 30% reduction in errors compared to models with one hidden layer. Errors were primarily caused by data imbalance and specific signal characteristics that were not correctly identified by the model. The causes of larger prediction errors were also analyzed, revealing that certain PPG signals exhibited features resembling those of older or younger age groups, which influenced the model’s errors. Further optimization of the model and data processing can significantly improve prediction accuracy, potentially making this approach an effective tool for real-time medical prediction.

KLJUČNE REČI / KEYWORDS:

PPG signals, machine learning, neural networks, age prediction, data processing, cardiovascular health

PROJEKAT / ACKNOWLEDGEMENT:

LITERATURA / REFERENCES:

  • ECG for Everybody smartphone app: Pulse HRV by Camera BLE ECG. Available: https://play.google.com/store/apps/details?id=srb.ctb.pulse.heartrate.ca mera.ecg4everybody.
  • J. Allen, „Photoplethysmography and its application in clinical physiological measurement,“Physiological Measurement, vol. 28, no. 3, pp. R1–R39, 2007.
  • M. Elgendi, „On the Analysis of Fingertip Photoplethysmogram Signals,“Current Cardiology Reviews, vol. 8, no. 1, pp. 14–25, 2012.
  • K. He, X. Zhang, S. Ren, and J. Sun, „Deep Residual Learning for Image Recognition,“ in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, 2016.
  • D. P. Kingma and J. Ba, „Adam: A Method for Stochastic Optimization,“arXiv preprint, arXiv:1412.6980, 2014.
  • Y. Liang,     Z.            Chen,     R.           Ward,     and         M.           Elgendi, „Photoplethysmography and Deep Learning: Enhancing PPG Signal Analysis,“Frontiers in Physiology, vol. 9, p. 1038, 2018.
  • J. and W. J. Tompkins, „A Real-Time QRS Detection Algorithm,“IEEE Transactions on Biomedical Engineering, vol. BME-32, no. 3, pp. 230–236, 1985.
  • S. Raj, W. Wang, and W. Zhao, „Photoplethysmography-based Estimation of Biological Age Using Machine Learning,“Scientific Reports, vol. 10, p. 6804, 2020.
  • R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, „Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization,“International Journal of Computer Vision, vol. 128, pp. 336–359, 2017.
  • Z. Zhang, „Photoplethysmography-Based Heart Rate Monitoring in Physical Activities via Joint Sparse Spectrum Reconstruction,“IEEE Transactions on Biomedical Engineering, vol. 62, no. 8, pp. 1902–1910, 2015.
  • M. A. F. Pimentel, D. A. Clifton, L. Clifton, and L. Tarassenko, „A Review of Novel Approaches to Vital Sign Monitoring in the Clinic and the Home,“Physiological Measurement, vol. 36, no. 7, pp. R1–R44, 2016.
  • J. Allen and A. Murray, „Age-Related Changes in the Characteristics of the Photoplethysmographic Pulse Shape at Various Body Sites,“Physiological Measurement, vol. 24, no. 2, pp. 297–307, 2004. Available: https://doi.org/10.1088/0967-3334/24/2/302.
  • Y. Zhang and Q. Yang, „A Survey on Multi-Task Learning,“IEEE Transactions on Knowledge and Data Engineering, vol. 34, no. 12, pp. 2309–2332, 2018.
  • Y. LeCun, Y. Bengio, and G. Hinton, „Deep Learning,“Nature, vol. 521, pp. 436–444, 2015.
  • F. Rundo, A. L. di Stallo, and C. Spampinato, „Advanced BioSignal Processing and Machine Learning Techniques for Cardiovascular Health Monitoring,“Biomedical Signal Processing and Control, vol. 55, p. 101641, 2019.
  • J. M. Mathias and S. Anagnostopoulou, „Automatic Age Estimation Using Deep Neural Networks,“Pattern Recognition Letters, vol. 125, pp. 82–91, 2019.
  • A. Joshi, A. Roy, and N. Sharma, „Transfer Learning for PPG Signal Analysis,“IEEE Access, vol. 6, pp. 47680–47691, 2018.
  • M. Elgendi, I. Norton, M. Brearley, D. Abbott, and D. Schuurmans, „Systolic Peak Detection in Acceleration Photoplethysmograms Measured from Emergency Responders in Tropical Conditions,“PLoS ONE, vol. 8, no. 10, p. e76585, 2013.
  • A. L. Goldberger, L. A. N. Amaral, L. Glass, J. M. Hausdorff, P. C. Ivanov, R. G. Mark, and H. E. Stanley, „Physiobank, Physiotoolkit, and Physionet: Components of a New Research Resource for Complex Physiologic Signals,“Circulation, vol. 101, no. 23, pp. e215–e220, 2000. Available: https://doi.org/10.1161/01.CIR.101.23.e215.
  • D. Yang, D. He, and W. Zhou, „Age Estimation Using Photoplethysmographic Signals and Deep Learning Techniques,“Journal of Medical Systems, vol. 42, p. 95, 2018.
  • J. M. Bland and D. G. Altman, „Statistical Methods for Assessing Agreement Between Two Methods of Clinical Measurement,“The Lancet, vol. 327, no. 8476, pp. 307–310, 1986.
  • C. G. Scully, J. Lee, J. Meyer, A. M. Gorbach, D. Granquist-Fraser, Y. Mendelson, and K. H. Chon, „Physiological Parameter Monitoring from Optical Recordings with a MobilePhone,“IEEE Transactions on Biomedical Engineering, vol. 59, no. 2, pp. 303–306, 2012.
  • M. Saeed, M. Villarroel, A. T. Reisner, G. Clifford, L. W. Lehman, G. Moody, T. Heldt, T. H. Kyaw, B. Moody, and R. G. Mark, „Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC-II): A Public-Access Intensive Care Unit Database,“Critical Care Medicine, vol. 39, no. 5, pp. 952–960, 2011.
  • D. J. McDuff, S. Gontarek, and R. W. Picard, „Improvements in Remote Cardiopulmonary Measurement Using a Five-Band Digital Camera,“IEEE Transactions on Biomedical Engineering, vol. 61, no. 10, pp. 2593–2601, 2014.
  • G. D. Clifford, F. Azuaje, and P. E. McSharry, Advanced Methods and Tools for ECG Data Analysis. Artech House Biomedical Engineering Series, 2006.
  • O. Yildirim, U. B. Baloglu, R. S. Tan, and U. R. Acharya, „A Deep Learning Model for Automated Arrhythmia Detection Using Photoplethysmography (PPG) Signals,“Computers in Biology and Medicine, vol. 113, p. 103387, 2019.